Do not enable calibration by default, so people get to see the warnings about codelets still not being calibrated

- create config files to store definitions
 - include/chameleon_config.h: definitions exposed to the public (installed file)
 - control/config.h: additional definitions used internally
- modify source code to call the config files when it is necessary (when the different "if defined" are used)

    - use (starpu_cpu_func_t) 1 trick, same as cuda_func
    - cpu funtions are not defined anymore avoiding the dependency to coreblas
    - add #if !defined(CHAMELEON_SIMULATION) where it is needed
    - remove dependency to the coreblas library (become useless)
    - remove useless simucblas, simulapacke libraries
    - remove CHAMELEON_SIMULATION_MAGMA cmake variable and definition
      - keep using CHAMELEON_USE_CUDA and CHAMELEON_USE_MAGMA to consider CUDA kernels
      - this avoid to introduce useless new variables
    - work on messages

don't force calibration in simulation mode. It does not make sense, and hides the StarPU detailed warning about missing calibration

Fix build with starpu1.1 which doesn't have starpu_mpi_comm_{rank,size}: let's just call RUNTIME_comm_{rank,size} which already do things properly

instead of introducing RUNTIME_distributed_barrier

Also make StarPU-MPI initialization assume that in the simgrid-mpi case we do not have any outer mpi initialization

MORSE_Distributed_size, MORSE_Distributed_rank so that applications do not
hardcode the use of MPI.

Introduce RUNTIME_distributed_rank, RUNTIME_distributed_size,
RUNTIME_distributed_barrier, so that MORSE does not hardcode the use of MPI
either.

This allows to use simgrid-mpi.

Add the possibility to enter Chameleon with an existing starpu (Need to be cleanup on the starpu_initialized function)

add MORSE_Pause/Resume function to avoid CPU consumption when no tasks have to be executed - call starpu_pause/resume, do nothing for quark (no need)

move init of starpu parameters in a single place (RUNTIME_init_scheduler) and use dmdas default scheduling strategy

change the way we include our own header files --> relative to the root - when plasma is in the same env, chameleon can take some headers not belonging to it (ex: #include descriptor.h, this file states in plasma install dir also) which make compilation errors

change copyright - correct whitespace - place cmake module depending on chameleon in cmake_modules and no more in cmake_modules/morse